1. Technical Field
The field of the invention relates generally to power production from distributed DC power sources, and more particularly to management of distributed DC power sources in series installations.
2. Description of Related Art
The recent increased interest in renewable energy has led to increased research in systems for distributed generation of energy, such as photovoltaic cells (PV), fuel cells, batteries (e.g., for hybrid cars), etc. Various topologies have been proposed for connecting these power sources to the load, taking into consideration various parameters, such as voltage/current requirements, operating conditions, reliability, safety, costs, etc. For example, most of these sources provide low voltage output (normally a few volts for one cell, or a few tens of volts for serially connected cells), so that many of them need to be connected serially to achieve the required operating voltage. Conversely, a serial connection may fail to provide the required current, so that several strings of serial connections may need to be connected in parallel to provide the required current.
It is also known that power generation from each of these sources depends on manufacturing, operating, and environmental conditions. For example, various inconsistencies in manufacturing may cause two identical sources to provide different output characteristics. Similarly, two identical sources may react differently to operating and/or environmental conditions, such as load, temperature, etc. In practical installations, different source may also experience different environmental conditions, e.g., in solar power installations some panels may be exposed to full sun, while others be shaded, thereby delivering different power output. In a multiple-battery installation, some of the batteries may age differently, thereby delivering different power output. While these problems and the solutions provided by the subject invention are applicable to any distributed power system, the following discussion turns to solar energy so as to provide better understanding by way of a concrete example.
A conventional installation of solar power system 10 is illustrated in FIG. 1. Since the voltage provided by each individual solar panel 101 is low, several panels are connected in series to form a string of panels 103. For a large installation, when higher current is required, several strings 103 may be connected in parallel to form the overall system 10. The solar panels are mounted outdoors, and their leads are connected to a maximum power point tracking (MPPT) module 107 and then to an inverter 104. The MPPT 107 is typically implemented as part of the inverter 104. The harvested power from the DC sources is delivered to the inverter 104, which converts the fluctuating direct-current (DC) into alternating-current (AC) having a desired voltage and frequency, which is usually 110V or 220V at 60 Hz, or 220V at 50 Hz (It is interesting to note the even in the US many inverters produce 220V, which is then split into two 110V feeds in the electric box). The AC current from the inverter 104 may then be used for operating electric appliances or fed to the power grid. Alternatively, if the installation is not tied to the grid, the power extracted from the inverter may be directed to a conversion and charge/discharge circuit to store the excess power created as charge in batteries. In case of a battery-tied application, the inversion stage might be skipped altogether, and the DC output of the MPPT stage 107 may be fed into the charge/discharge circuit.
As noted above, each solar panel 101 supplies relatively very low voltage and current. The problem facing the solar array designer is to produce a standard AC current at 120V or 220V root-mean-square (RMS) from a combination of the low voltages of the solar panels. The delivery of high power from a low voltage requires very high currents, which cause large conduction losses on the order of the second power of the current (I2). Furthermore, a power inverter, such as the inverter 104, which is used to convert DC current to AC current, is most efficient when its input voltage is slightly higher than its output RMS voltage multiplied by the square root of 2. Hence, in many applications, the power sources, such as the solar panels 101, are combined in order to reach the correct voltage or current. The most common method connects the power sources in series in order to reach the desirable voltage and in parallel in order to reach the desirable current, as shown in FIG. 1. A large number of the panels 101 are connected into a string 103 and the strings 103 are connected in parallel to the power inverter 104. The panels 101 are connected in series in order to reach the minimal voltage required for the inverter. Multiple strings 103 arc connected in parallel into an array to supply higher current, so as to enable higher power output.
While this configuration is advantageous in terms of cost and architecture simplicity, several drawbacks have been identified in the literature for such architecture. One recognized drawback is inefficiencies cause by non-optimal power draw from each individual panel, as explained below. As explained above, the output of the DC power sources is influenced by many conditions. Therefore, to maximize the power draw from each source, one needs to draw the combination of voltage and current that provides the peak power for the currently prevailing conditions. As conditions change, the combination of voltage and current draw may need to be changed as well.
FIG. 2 illustrates one serial string of DC sources, e.g., solar panels 201 a-201 d, connected to MPPT circuit 207 and inverter 204. The current versus voltage (IV) characteristics plotted (210a-210d) to the left of each DC source 201. For each DC source 201, the current decreases as the output voltage increases. At some voltage value the current goes to zero, and in some applications may assume a negative value, meaning that the source becomes a sink. Bypass diodes are used to prevent the source from becoming a sink. The power output of each source 201, which is equal to the product of current and voltage (P=I*V), varies depending on the voltage drawn from the source. At a certain current and voltage, close to the falling off point of the current, the power reaches its maximum. It is desirable to operate a power generating cell at this maximum power point. The purpose of the MPPT is to find this point and operate the system at this point so as to draw the maximum power from the sources.
In a typical, conventional solar panel array, different algorithms and techniques are used to optimize the integrated power output of the system 10 using the MPPT module 107. The MPPT module 107 receives the current extracted from all of the solar panels together and tracks the maximum power point for this current to provide the maximum average power such that if more current is extracted, the average voltage from the panels starts to drop, thus lowering the harvested power. The MPPT module 107 maintains a current that yields the maximum average power from the overall system 10.
However, since the sources 201a-201d are connected in series to a single MPPT 207, the MPPT must select a single point, which would be somewhat of an average of the MPP of the serially connected sources. In practice, it is very likely that the MPPT would operate at an I-V point that is optimum to only a few or none of the sources. In the example of FIG. 2, the selected point is the maximum power point for source 201b, but is off the maximum power point for sources 201a, 201c and 201d. Consequently, the arrangement is not operated at best achievable efficiency.
Turning back to the example of a solar system 10 of FIG. 1, fixing a predetermined constant output voltage from the strings 103 may cause the solar panels to supply lower output power than otherwise possible. Further, each string carries a single current that is passed through all of the solar panels along the string. If the solar panels are mismatched due to manufacturing differences, aging or if they malfunction or are placed under different shading conditions, the current, voltage and power output of each panel will be different. Forcing a single current through all of the panels of the string causes the individual panels to work at a non-optimal power point and can also cause panels which are highly mismatched to generate “hot spots” due to the high current flowing through them. Due to these and other drawbacks of conventional centralized methods, the solar panels have to be matched properly. In some cases external diodes are used to bypass the panels that are highly mismatched. In conventional multiple string configurations all strings have to be composed of exactly the same number of solar panels and the panels are selected of the same model and must be install at exactly the same spatial orientation, being exposed to the same sunlight conditions at all times. This is difficult to achieve and can be very costly.